The present invention relates to the identification of compounds that are suitable for imaging amyloid deposits in living patients. More specifically, the present invention relates to a method of imaging amyloid deposits in brain in vivo to allow antemortem diagnosis of Alzheimer's Disease. The present invention also relates to therapeutic uses for such compounds.
Alzheimer's Disease ("AD") is a neurodegenerative illness characterized by memory loss and other cognitive deficits. McKhann et al., Neurology 34: 939 (1984). It is the most common cause of dementia in the United States. AD can strike persons as young as 40-50 years of age. Yet, because the presence of the disease is difficult to determine without dangerous brain biopsy, the time of onset is unknown. The prevalence of AD increases with age, with estimates of the affected population reaching as high as 40-50% by ages 85-90. Evans et al., JAMA 262: 2551 (1989); Katzman, Neurology 43: 13 (1993).
By definition, AD is definitively diagnosed through examination of brain tissue, usually at autopsy. Khachaturian, Arch. Neurol. 42: 1097 (1985); McKhann et al., Neurology 34: 939 (1984). Neuropathologically, this disease is characterized by the presence of neuritic plaques (NP), neurofibrillary tangles (NFT), and neuronal loss, along with a variety of other findings. Mann, Mech. Ageing Dev. 31: 213 (1985). Post-mortem slices of brain tissue of victims of Alzheimer's disease exhibit the presence of amyloid in the form of proteinaceous extracellular cores of the neuritic plaques that are characteristic of AD.
The amyloid cores of these neuritic plaques are composed of a protein called the .beta.-amyloid (A.beta.) that is arranged in a predominately beta-pleated sheet configuration. Mori et al., Journal of Biological Chemistry 267: 17082 (1992); Kirschner et al., PNAS 83: 503 (1986). Neuritic plaques are an early and invariant aspect of the disease. Mann et al., J. Neurol. Sci. 89: 169; Mann, Mech. Ageing Dev. 31: 213 (1985); Terry et al., J. Neuropathol. Exp. Neurol 46: 262 (1987).
The initial deposition of A.beta. probably occurs long before clinical symptoms are noticeable. The currently recommended "minimum microscopic criteria" for the diagnosis of AD is based on the number of neuritic plaques found in brain. Khachaturian, Arch. Neurol., supra (1985). Unfortunately, assessment of neuritic plaque counts must be delayed until after death.
Amyloid-containing neuritic plaques are a prominent feature of selective areas of the brain in AD as well as Downs Syndrome and in persons homozygous for the apolipoprotein E4 allele who are very likely to develop AD. Corder et al., Science 261: 921 (1993); Divry, P., J. Neurol. Psych. 27: 643-657 (1927); Wisniewski et al., in Zimmerman, H. M. (ed.): PROGRESS IN NEUROPATHOLOGY, (Grune and Stratton, N.Y. 1973) pp. 1-26. Brain amyloid is readily demonstrated by staining brain sections with thioflavin S or Congo red. Puchtler et al., J. Histochem. Cytochem. 10: 35 (1962). Congo red stained amyloid is characterized by a dichroic appearance, exhibiting a yellow-green polarization color. The dichroic binding is the result of the beta-pleated sheet structure of the amyloid proteins. Glenner, G. N. Eng. J. Med. 302: 1283 (1980). A detailed discussion of the biochemistry and histochemistry of amyloid can be found in Glenner, N. Eng. J. Med., 302: 1333 (1980).
Thus far, diagnosis of AD has been achieved mostly through clinical criteria evaluation, brain biopsies and post mortem tissue studies. Research efforts to develop methods for diagnosing Alzheimer's disease in vivo include (1) genetic testing, (2) immunoassay methods and (3) imaging techniques.
Evidence that abnormalities in A.beta. metabolism are necessary and sufficient for the development of AD is based on the discovery of point mutations in the A.beta. precursor protein in several rare families with an autosomal dominant form of AD. Hardy, Nature Genetics 1: 233 (1992); Hardy et al., Science 256: 184 (1992). These mutations occur near the N- and C-terminal cleavage points necessary for the generation of A.beta. from its precursor protein. St. George-Hyslop et al., Science 235: 885 (1987); Kang et al., Nature 325: 733 (1987); Potter WO 92/17152. Genetic analysis of a large number of AD families has demonstrated, however, that AD is genetically heterogeneous. St. George-Hyslop et al., Nature 347: 194 (1990). Linkage to chromosome 21 markers is shown in only some families with early-onset AD and in no families with late-onset AD. More recently a gene on chromosome 14 whose product is predicted to contain multiple transmembrane domains and resembles an integral membrane protein has been identified by Sherrington et al., Nature 375: 754-760 (1995 ). This gene may account for up to 70% of early-onset autosomal dominant AD. Preliminary data suggests that this chromosome 14 mutation causes an increase in the production of A.beta.. Scheuner et al., Soc. Neurosci. Abstr. 21: 1500 (1995). A mutation on a very similar gene has been identified on chromosome 1 in Volga German kindreds with early-onset AD. Levy-Lahad et al., Science 269: 973-977 (1995).
Screening for apolipoprotein E genotype has been suggested as an aid in the diagnosis of AD. Scott, Nature 366: 502 (1993); Roses, Ann. Neurol. 38: 6-14 (1995). Difficulties arise with this technology, however, because the apolipoprotein E4 allele is only a risk factor for AD, not a disease marker. It is absent in many AD patients and present in many non-demented elderly people. Bird, Ann. Neurol. 38: 2-4 (1995).
Immunoassay methods have been developed for detecting the presence of neurochemical markers in AD patients and to detect an AD related amyloid protein in cerebral spinal fluid. Warner, Anal. Chem. 59: 1203A (1987); World Patent No. 92/17152 by Potter; Glenner et al., U.S. Pat. No. 4,666,829. These methods for diagnosing AD have not been proven to detect AD in all patients, particularly at early stages of the disease and are relatively invasive, requiring a spinal tap. Also, attempts have been made to develop monoclonal antibodies as probes for imaging of A.beta.. Majocha et al., J. Nucl. Med., 33: 2184 (1992); Majocha et al., WO 89/06242 and Majocha et al., U.S. Pat. No. 5,231,000. The major disadvantage of antibody probes is the difficulty in getting these large molecules across the blood-brain barrier. Using antibodies for in vivo diagnosis of AD would require marked abnormalities in the blood-brain barrier in order to gain access into the brain. There is no convincing functional evidence that abnormalities in the blood-brain barrier reliably exist in AD. Kalaria, Cerebrovascular & Brain Metabolism Reviews 4: 226 (1992).
A.beta. antibodies are also disadvantageous for use in AD diagnostics because they typically stain deposits of A.beta. containing non-.beta.-sheet (non-fibrillar) A.beta. in addition to the neuritic plaques. Yamaguchi et al., Acta Neuropathol., 77: 314 (1989). These deposits may be a separate type of lesion, not necessarily involved in the dementing process of AD. The latter is suggested by findings of nonfibrillar amyloid deposits in cognitively normal controls and aged dogs. Moran et al., Medicina Clinica 98: 19 (1992); Shimada et al., Journal of Veterinary Medical Science 54: 137 (1992); Ishihara et al., Brain Res. 548: 196 (1991); Giaccone et al., Neurosci. Lett. 114: 178 (1990). Even if non-fibrillar amyloid deposits are forerunners of neuritic plaques, the key pathological event in AD may be the process that turns the apparently benign non-fibrillar amyloid deposit into the neuritic plaque with its associated halo of degeneration. Therefore, a probe is needed that is specific for the fibrillar A.beta. deposits and NFTs as a more specific marker for AD pathophysiology than antibodies that would also label non-fibrillar amyloid deposits.
Recently, radiolabeled A.beta. peptide has been used to label diffuse, compact and neuritic type plaques in sections of AD brain. Maggio et al., WO 93/04194. However, these peptides share all of the disadvantages of antibodies. Specifically, peptides do not normally cross the blood-brain barrier in amounts necessary for imaging.
Congo red may be used for diagnosing amyloidosis in vivo in non-brain parenchymal tissues. However, Congo red is probably not suitable for in vivo diagnosis of A.beta. deposits in brain because only 0.03% of an injected dose of iodinated Congo red can enter the brain parenchyma. Tubis et al., J. Amer. Pharm. Assn. 49: 422 (1960). Radioiodinated bisdiazobenzidine compounds related to Congo red, such as Benzo Orange R and Direct Blue 4, have been proposed to be useful in vitro and in vivo to detect the presence and location of amyloid deposits in an organ of a patient. Quay et al., U.S. Pat. Nos. 5,039,511 and 4,933,156. However, like Congo red, all of the compounds proposed by Quay contain strongly acidic sulfonic acid groups which severely limit entry of these compounds into the brain making it extremely difficult to attain an imaging effective quantity or detectable quantity in the brain parenchyma.
The inability to assess amyloid deposition in AD until after death impedes the study of this devastating illness. A method of quantifying amyloid deposition before death is needed both as a diagnostic tool in mild or clinically confusing cases as well as in monitoring the effectiveness of therapies targeted at preventing A.beta. deposition. Therefore, it remains of utmost importance to develop a safe and specific method for diagnosing AD before death by imaging amyloid in brain parenchyma in vivo. Even though various attempts have been made to diagnose AD in vivo, currently, there are no antemortem probes for brain amyloid. No method has utilized a high affinity probe for amyloid that has low toxicity, can cross the blood-brain barrier, and binds more effectively to AD brain than to normal brain in order to identify AD amyloid deposits in brain before a patient's death. Thus, no in vivo method for AD diagnosis has been demonstrated to meet these criteria.
Very recent data suggest that amyloid-binding compounds will have therapeutic potential in AD and type 2 diabetes mellitus. As mentioned above, there are two broad categories of plaques in AD brain, diffuse and neuritic (classical). Diffuse plaques do not appear to induce morphological reactions such as the reactive astrocytes, dystrophic neurites, microglia cells, synapse loss, and full complement activation found in neuritic plaques. Joachim et al., Am. J. Pathol. 135: 309 (1989); Masliah et al., loc. cit. 137: 1293 (1990); Lue and Rogers, Dementia 3: 308 (1992). These morphological reactions all signify that neurotoxic and cell degenerative processes are occurring in the areas adjacent to the fibrillar A.beta. deposits of neuritic plaques. A.beta.-induced neurotoxicity and cell degeneration has been reported in a number of cell types in vitro. Yankner et al., Science 250: 279 (1990); Roher et al., BBRC 174: 572 (1991); Frautschy et al., Proc. Natl. Acad. Sci. 88: 83362 (1991); Shearman et al., loc. cit. 91: 1470 (1994). It has been shown that aggregation of the A.beta. peptide is necessary for in vitro neurotoxicity. Yankner, Neurobiol. Aging 13: 615 (1992). Differences in the state of aggregation of A.beta. in diffuse and neuritic plaques may explain the lack of neurotoxic response surrounding the diffuse plaque. Lorenzo and Yankner, Proc. Natl. Acad. Sci., 91: 12243 (1994). Recently, three laboratories have reported results which suggest that Congo red inhibits A.beta.-induced neurotoxicity and cell degeneration in vitro. Burgevin et al., NeuroReport 5: 2429 (1994); Lorenzo and Yankner, Proc. Natl. Acad. Sci. 91: 12243 (1994); Pollack et al., Neuroscience Letters 184: 113 (1995); Pollack et al., Neuroscience Letters 197: 211 (1995). The mechanism appears to involve both inhibition of fibril formation and prevention of the neurotoxic properties of formed fibrils. Lorenzo and Yankner, Proc. Natl. Acad. Sci. 91: 12243 (1994). Congo red also has been shown to protect pancreatic islet cells from the toxicity caused by amylin. Lorenzo and Yankner, Proc. Natl. Acad. Sci. 91: 12243 (1994). Amylin is a fibrillar peptide similar to A.beta. which accumulates in the pancreas in type 2 diabetes mellitus.
It is known in the art that certain azo dyes may be carcinogenic. Morgan et al. Environmental Health Perspectives, 102 (supp.) 2: 63-78, (1994). This potential carcinogenicity appears to be based largely on the fact that azo dyes are extensively metabolized to the free parent amine by intestinal bacteria. Cerniglia et al., Biochem. Biophys. Res. Com., 107: 1224-1229, (1982). In the case of benzidine dyes (and many other substituted benzidines), it is the free amine which is the carcinogen. These facts have little implication for amyloid imaging studies in which an extremely minute amount of the high specific activity radiolabelled dye would be directly injected into the blood stream. In this case, the amount administered would be negligible and the dye would by-pass the intestinal bacteria.
In the case of therapeutic usage, these facts have critical importance. Release of a known carcinogen from a therapeutic compound is unacceptable. A second problem with diazo dye metabolism is that much of the administered drug is metabolized by intestinal bacteria prior to absorption. This lowered bioavailability remains a disadvantage even if the metabolites released are innocuous.
Thus, a need exists for amyloid binding compounds which are similar to Congo red but which enter the brain (Congo Red does not). Such compounds could be used in preventing cell degeneration associated with fibril formation and thereby treat pathological conditions in amyloid associated diseases, such as AD and Downs Syndrome and in treating pancreatic islet cell toxicity in type 2 diabetes mellitus.
A further need exists for amyloid binding compounds that are non-toxic and bioavailable and, consequently, can be used in therapeutics.